BmILF and i-motif structure are involved in transcriptional regulation of BmPOUM2 in Bombyx mori

Abstract

Guanine-rich and cytosine-rich DNA can form four-stranded DNA secondary structures called G-quadruplex (G4) and i-motif, respectively. These structures widely exist in genomes and play important roles in transcription, replication, translation and protection of telomeres. In this study, G4 and i-motif structures were identified in the promoter of the transcription factor gene BmPOUM2, which regulates the expression of the wing disc cuticle protein gene (BmWCP4) during metamorphosis. Disruption of the i-motif structure by base mutation, anti-sense oligonucleotides (ASOs) or inhibitory ligands resulted in significant decrease in the activity of the BmPOUM2 promoter. A novel i-motif binding protein (BmILF) was identified by pull-down experiment. BmILF specifically bound to the i-motif and activated the transcription of BmPOUM2. The promoter activity of BmPOUM2 was enhanced when BmILF was over-expressed and decreased when BmILF was knocked-down by RNA interference. This study for the first time demonstrated that BmILF and the i-motif structure participated in the regulation of gene transcription in insect metamorphosis and provides new insights into the molecular mechanism of the secondary structures in epigenetic regulation of gene transcription.

INTRODUCTION

Transcriptional factors of the POU domain family, including Pit, Oct, and Unc, have been identified from many species and reported to play critical roles in many developmental processes (1–3). In mammals, POU transcription factors regulate the development of the neuroendocrine systems (4). In insects, POU transcription factor plays a critical role in the neuronal lineage and wiring (5), cell proliferation and differentiation in the developing wing imaginal disc (6), regulation of sericin-1 gene (7), fibroin gene (8) and dopa decarboxylase (DDC) gene (9). In Bombyx mori, an important economic insect as well as a model lepidopteran insect, POUM was reported to regulate the transcription of sericin-1 gene (7) and to regulate the expression of B. mori diapause hormone and pheromone biosynthesis activating neuropeptide (DH-PBAN) gene during metamorphosis (10). In our previous study, BmPOUM2 was shown to regulate the expression of wing disc cuticle protein gene (BmWCP4) during metamorphosis of B. mori. Down regulation of BmPOUM2 expression resulted in failure to complete metamorphosis accompanied with wing disc growth and differentiation arrest and failure for larvae to spin cocoons (11,12). Study of the regulatory mechanism of BmPOUM2 would be critical to explain wing development.

G-quadruplex (G4) and i-motif have been identified as DNA secondary structures. The G4 structure is formed by stacked G-tetrads, which are square co-planar arrays of four guanine bases and stabilized by monovalent cation such as K⁺ or Na⁺ (13–16). G4 has been found to regulate multiple biological processes such as replication, transcription and telomeres protection (17,18). More than 700 orthologous genes were found to have the probability to form G4 in human, chimpanzee, mouse and rat, suggesting the widespread regulatory influence of G4 motifs (19). In human, the regulatory regions of >40% genes were predicted to have G4 structure and to be involved in the gene regulation at the level of transcription (20). Stable G4 structure identified in the promoter region of the WNT1 gene suppressed its expression and its downstream signaling pathways, resulting in the inhibition of Wnt1 mediated migration and invasion activities of cancer cells (21). Similar effect was also demonstrated in the promoter of c-MYC (22,23). In contrast, G4 located the upstream transcriptional starting site (TSS) of col2al, fdz5 and nog3 act as an enhancer for transcription (24).

The complementary strand of G-rich sequence is unsurprisingly C-rich and it can form another DNA secondary structure called i-motif. I-motif is a four-stranded structure formed by intercalated hemiprotonated cytosine–cytosine (C-C⁺) base pairs under acidic condition (25) or at neutral pH by molecular crowding of the cosolutes (26) or in high pressure (27). Unlike G4 structure, studies on the i-motif structure and its function are limited. I-motif structure has been found in the promoter of genes BCL2 (28), Th (29) and KRAS (30) and has been identified as a recognition binding site for transcription factors to activate gene transcription (31).

In this study, we identified G4 and i-motif secondary structures in the promoter region of the transcription factor BmPOUM2 and demonstrated for the first time that a nuclear transcription factor protein, BmILF, bound to the i-motif structure, regulating the transcription of BmPOUM2 in B. mori.

MATERIALS AND METHODS

Insect and cell line

Bombyx mori strain Dazao was provided by the Research and Development Center of the Sericultural Research Institute of the Academy of Agricultural Sciences of Guangdong Province, China. Larvae were raised on fresh mulberry leaves at 26°C under a 14/10 h light/dark photoperiod.

A B. mori cell line DZNU-Bm-12 (Bm12) originally developed from ovarian tissues (32) was maintained at 28°C in Grace medium (Invitrogen, USA) supplemented with 10% FBS.

Cloning of the BmPOUM2 promoter

Genomic DNA was prepared from fifth instar larvae of B. mori using the phenol extraction method (33). The primers used for amplifying the promoter of BmPOUM2 from genomic DNA were: Forward primer: 5′-TGATGACACGACACAATAACC-3′; Reverse primer: 5′-GGCAGGCTCAGCACTACG-3′. Genomic DNA was denatured at 94°C for 3 min, followed by a 30-cycle reaction (94°C for 30 s; 55°C for 30 s; 72°C for 2 min). The primers used for amplifying the promoter of BmPOUM2 from 1638-pMD-18T vector were: Forward primer: 5′-GTTTATTAAATTGGGGCCGTGA-3′; Reverse primer: 5′-ACGGCAGCACGGAGCGA-3′. Template DNA was denatured at 94°C for 3 min, followed by 27 cycles of 94°C for 30 s, 55°C for 30 s and 72°C for 30 s for each cycle. In this study, different doses of dimethyl sulphoxide (DMSO) were added into the PCR reaction mix or the PCR reaction conditions were adjusted to obtain the target products. PCR products were separated on 1% agarose gels for verification.

Circular dichroism (CD) analysis

CD experiments were performed with a J-815 CD spectrometer (Jasco International, USA). All spectra were collected with a wavelength range from 220 to 350 nm, 1 nm step width and 1 s response time. The CD spectra are representations of three averaged scans of the same sample taken at room temperature and are baseline corrected for signal contributions due to the buffer. DNA oligonucleotide sequences (5 μM) used for analyzing G4 structure were heated at 95°C for 10 min in 50 mM Tris buffer (pH 7.5) with or without 100 mM KCl, and slowly cooled to room temperature over 4 h period. DNA oligonucleotide sequences (5 μM) used for analyzing i-motif structure were heated at 95°C for 10 min in 50 mM Tris–acetate buffer at different pHs, and slowly cooled to room temperature over 4 h period. For the incubation with TMPyP2 or TMPyP4 (Frontier Scientific, USA), TMPyP2 or TMPyP4 (dissolved in ddH₂O) was added into the annealed mix at a final concentration of 25, 50, 100 μM and incubated overnight at 4°C. The CD melting curves were obtained at 286 nm between 20 and 95°C with a heating rate of 1°C/min. The synthesized ssDNA oligonucleotides that contained the i-motif region (5 μM) were heated at 95°C for 10 min in 50 mM Tris-acetate buffer at pH 4.1, and slowly cooled to room temperature over 4 h period to allow i-motif structure to form. TMPyP2 or TMPyP4 was then added into the solution at a final concentration of 25 μM and incubated overnight at 4°C, followed by CD analysis.

Luciferase assay for promoter activity analysis

Bm12 cells were inoculated into culture media in 12- or 24-well culture plates (Corning, USA) and cultured for 12 h. Cell transfection and co-transfection were conducted when the cell density achieved 80%. To normalize the firefly luciferase activity, the renilla luciferase vector pRL-SV40 (Promega, USA) was co-transfected with each of the pGL3-derived vectors. For transfection, a mixture of 50 μl containing 200 ng of the wild-type or mutated 1360-BmPOUM2-P-pGL3 reporter plasmid DNA, 50 ng internal control plasmid and 2 μl Fugene HD transfection reagent (Promega, USA) in the Opti-MEM Reduced Serum Medium (Life Technologies, USA) was added to Bm12 cells in Grace medium with or without TMPyP2 or TMPyP4. The cells were cultured for 48 h at 28°C before the promoter activity assay.

For co-transfection, a mixture of 50 μl containing 200 ng of the wild-type or mutated 1360-BmPOUM2-P-pGL3 reporter plasmid DNA, 1 μg EGFP or BmILF-EGFP plasmid DNA, 200 ng internal control plasmid (pRL-SV40 vector) and 6 μl Fugene HD transfection reagent in the Opti-MEM Reduced Serum Medium was added to the cells in Grace medium. The cells were cultured for 48 h at 28°C before the promoter activity assay.

For the anti-sense oligonucleotides (ASOs) disturbing assays, a mixture of 50 μl containing 200 ng of the wild-type or mutated 1360-BmPOUM2-P-pGL3 reporter plasmid DNA, 2 μl ASOs (100 μM), 300 ng internal control plasmid (pRL-SV40 vector) and 6 μl Fugene HD transfection reagent in the Opti-MEM Reduced Serum Medium was added to the cells in Grace medium in each well. The cells were cultured for 48 h at 28°C before the determination of promoter activity. The oligonucleotides used in the ASOs disturbing assays were: wild-type ASO-i-motif: 5′-GGGGCGCGAGGG-3′; mutant ASO-i-motif: 5′-GATGCGCGAGTG-3′; wild-type ASO-G4: 5′-CCCTCGCGCCCC-3′; mutant ASO-G4: 5′-CACTCGCGCATC-3′; non-specific-ssDNA: 5′-ATCGATCGATCG-3′; F-ASO-β-actin: 5′-TTTCGCGTGTAT-3′; R-ASO-β-actin: 5′-ATACACGCGAAA-3′.

The measurement of luciferase activity was carried out as described previously (11). Briefly, the transfected cells were washed twice with PBS and lysed in 150 μl Passive Lysis Buffer (Promega, USA). Samples were centrifuged at 1000 g for 5 min at room temperature. The supernatant was used to analyze the luciferase activity using the Dual-Luciferase Assay System according to the manufacturer's protocol with a luminometer (IBA7300, Veritas, Turner Biosystems). The luciferase activity was normalized to the renilla luciferase activity. All assays were repeated at least three times. The luciferase activity was represented as mean ± standard error (SE). Statistical significance of the luciferase activity was analyzed using Student's t-test.

Protein pull-down and liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis

Protein pull-down assays referenced the previous study (31). Nuclear proteins of Bm12 cells were extracted according to the instructions of NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific, USA). The protein concentration was determined using the BCA Protein Assay Reagent Kit (Thermo Scientific, USA). DNA oligonucleotide sequences of G4 structure were heated at 95°C for 10 min in 50 mM Tris buffer at pH 7.5 and slowly cooled to room temperature over 4 h. DNA oligonucleotide sequences of i-motif structure were heated at 95°C for 10 min in 50 mM Tris-acetate buffer at pH 4.1 and slowly cooled to room temperature over 4 h. Single-stranded biotinylated DNA oligonucleotide (20 μg) was incubated with 100 μg streptavidin-coated Dynabeads (Life Technologies, USA) in 400 μl binding buffer (10 mM Tris, pH 7.5, 1 mM EDTA, 1 M NaCl, 0.003% NP40) for 30 min at room temperature with constant and slow rotation. After twice washing with binding buffer, the immobilized DNA was incubated for 30 min in 400 μl blocking buffer (2.5 mg/ml BSA, 10 mM HEPES, pH 7.6, 100 mM potassium glutamate, 2.5 mM DTT, 10 mM magnesium acetate, 5 mM EGTA, 3.5% glycerol with 0.003% NP40 and 5 mg/ml polyvinylpyrrolidone). Then 100 μg nuclear proteins was incubated with the immobilized DNA for 4 h at 4°C in 400 μl protein binding buffer (10 mM HEPES, pH 7.6, 100 mM potassium glutamate, 80 mM KCl, 2.5 mM DTT, 10 mM magnesium acetate, 5 mM EGTA, 3.5% glycerol with 0.001% NP40 and 1 μg non-specific DNA) with constant and slow rotation. After 4 h, the DNA/protein complexes were washed three times with 400 μl washing buffer (10 mM HEPES, pH 7.6, 100 mM potassium glutamate, 2.5 mM DTT, 10 mM magnesium acetate, 5 mM EGTA, 3.5% glycerol, 0.5 mg/ml BSA, 0.05% NP40). Then the proteins bound to the DNA were eluted in 20 μl SDS-PAGE sample buffer (50 mM Tris, 100 mM DTT, 2% SDS, 0.1% Bromophenol blue, 10% glycerol). The eluted proteins were subjected to 12% SDS-PAGE. The gels were stained with Coomassie Brilliant Blue R-250 for more than 4 h and distained with distaining solution (10% acetic acid, 5% ethanol, 85% water). The differentiated protein bands were excised and analyzed for protein identification in Huijun Biotechnology (Guangzhou, China) using liquid chromatography-tandem mass spectrometry (LC-MS/MS).

Expression and purification of recombinant BmILF protein

Total RNA extraction from Bm12 cells and reverse transcription were conducted using the same method as described previously (11). BmILF open reading frame (ORF) fragment was amplified using cDNA cloned from Bm12 cells with forward primer: 5′-GAATTCATGGTGCGCGGCATC-3′ and reverse primer: 5′-AAGCTTTTACACT GAATCTTTGTCGTCGGT-3′. Obtained DNA fragment was then sub-cloned into the pET-28a vector in fusion with a 6 × His tag between EcoR I and Hind III restriction enzyme sites, generating a BmILF-pET-28a recombinant expression vector.

The recombinant BmILF protein was expressed in Escherichia coli cells (BL21), which were grown in Luria-Bertani (LB) medium containing 50 μg/ml kanamycin at 18°C with 12 h induction of 0.1 mM Isopropyl-β-d-thiogalactopyranoside (IPTG). The cells were collected by centrifugation and re-suspended in the binding buffer (0.5 M NaCl, 20 mM Tris–HCl, 5 mM Imidazole, pH 7.9 and 1 mM PMSF). The suspension was centrifuged at 10 000 g at 4°C for 5 min after being lysed by sonication, and then purified with Ni-chelating affinity chromatography using the His-Bind^® Kit according to the manufacturer's protocol (Novagen, USA). The protein concentration was determined using the BCA Protein Assay Reagent Kit (Thermo Scientific, USA).

Electrophoretic mobility shift assay (EMSA) for protein-DNA binding

EMSA was conducted using the Light Shift Chemiluminescent EMSA Kit (Thermo Scientific, USA). The wild-type and mutant oligonucleotides were labeled with biotin at the 5′ end and heated at 95°C for 10 min in 50 mM Tris-acetate buffer at pH 4.1 and slowly cooled to room temperature. The biotin labeled oligonucleotides at the 5′ end was synthesized in Invitrogen (Shanghai, China).

Binding reactions were performed according to the instruction of the EMSA Kit. Briefly, reactions were conducted in a 20 μl mixture containing (1 × binding buffer, 2.5% glycerol, 0.05% NP-40, 5 mM MgCl₂, 4 mM EDTA, recombinant protein BmILF and 20 fmol of a biotinylated end-labeled probe) at room temperature for 20 min. For the competition assay, cold probes (un-biotinylated) were added to the binding reaction. The samples were then separated in a 6% or 12% polyacrylamide gel on the ice at 100 V for 1.5 or 3 h. After electrophoresis, the gel was blotted onto a Nylonpositively charged membrane (Amersham Biosciences, USA). The membranes were then developed using the Light Shift Chemiluminescent EMSA Kit according to the manufacturer's protocol. The oligonucleotide probes used were showed as followed: Wild-type probe: 5′-TTGTTGCCCCGCCCCTCGGCCCCCTCGCGCCCCGCAC TGG-3′; Mutant probe (the mutant sites were underlined): 5′-TTGTTGC ACTGCTACTCGGCTCACTCGCGCATCGCACTGG-3′.

Chromatin Immunoprecipitation (ChIP) assay

ChIP assay was conducted to detect the binding of BmILF with the i-motif of the BmPOUM2 promoter in Bm12 cells. Cells were crosslinked with 1% formaldehyde for 10 min at room temperature after being transfected with BmILF-EGFP or EGFP (control) plasmids for 48 h. Glycine was added to terminate the fixation and the cells were washed twice with one media volume of ice-cold PBS. Cells in 1 ml of ice-cold PBS and 10 μl Halt Cocktail were collected by scraping. The cells were centrifuged at 3000 g for 5 min and the cell pellet was broken up with extraction buffer containing protease/phosphatase inhibitors. Nuclei were collected by centrifuged at 9000 g for 3 min and digested by MNase. Digested chromatin was obtained by sonicating on ice with several pulses to break nuclear membrane and incubated for 20 s on ice between pulses and then centrifuged at 9000 g for 5 min. Immunoprecipitation experiment was performed following the manufacturer's instruction of Pierce™ Magnetic ChIP Kit (Thermo Scientific, USA). Ten micrograms of either rabbit anti-EGFP antibody or normal rabbit IgG (Thermo Scientific, USA) were used and IP reactions were incubated overnight at 4°C with constant mixing. DNA/protein/antibody complex was purified by incubating with ChIP Grade Protein A/G Magnetic Beads for 2 h at 4°C with mixing. Immunoprecipitated genomic DNA fragments were amplified by RT-PCR and qRT-PCR with primers: forward primer: 5′-GCGAGACGTACAGACGCGAG-3′; reverse primer: 5′-GCCCGCTAATGACGGTGTAA-3′. The length of the product is 86 bp. The specificity of primers was examined using Primer-blast (http://www.ncbi.nlm.nih.gov/tools/primer-blast/). For RT-PCR, template DNA was denatured at 96°C for 3 min, followed by 26 cycles of 96°C for 30 s, 55°C for 30 s and 72°C for 10 s for each cycle. For qRT-PCR, the SYBR Green Kit was used according to the manufacturer's instruction (TaKaRa, Japan). The qRT-PCR was performed at the following conditions: SYBR Premix Ex Taq (2×): 10 μl in 20 μl reaction volume; the primers concentrations: 0.4 μl (10 μM); the immunoprecipitated DNA samples: 4 μl. The mixtures were incubated at 95°C for 10 s, then 40 cycles at 95°C for 5 s, 60°C for 31 s using ABI7300 fluorescence quantitative PCR system. The enrichment of promoter sequence in immunoprecipitated DNA samples was normalized with DNA present in the 10% input material analyzed using the 2^−△△Ct method (34). Data is means ± SEM (N = 3). ***P < 0.001 (Student's t-test). The PCR products of the enriched promoter were sequenced.

RNA interference (RNAi)

For RNAi of BmILF in Bm12 cell line, a 458 bp unique fragment from 198–655 nts in the ORF of BmILF was chosen as a template for synthesizing gene-specific dsRNA. BmILF dsRNA was synthesized from the linearized template by using the T7 RiboMAXTM Express RNAi System (Promega, USA). DsRNA (4 μg) was transfected into Bm12 cells in Grace medium with 6 μl Fugene HD transfection reagent in the Opti-MEM Reduced Serum Medium. The cells were collected 48 and 72 h after transfection. Total RNA extraction and reverse transcription were conducted as above mentioned (11). RT-PCR was performed using specific primers designed based on the BmILF cDNA to detect the knock-down efficiency. The sequence of primers were: forward primer: 5′-ATGGTGCGCGGCATC-3′, reverse primer: 5′-TTACACTGAATCTTTGTCG TCGGT-3′. Template DNA was denatured at 94°C for 3 min, followed by 29 cycles of 94°C for 30 s, 55°C for 30 s and 72°C for 30 s for each cycle. To detect the effect of BmILF on BmPOUM2 promoter activity when BmILF was knocked down, a 50 μl mixture containing 4 μg BmILF dsRNA, 200 ng of the wild-type or mutated 1360-BmPOUM2-P-pGL3 reporter plasmid DNA, 200 ng internal control plasmid (pRL-SV40 vector, Promega) and 6 μl Fugene HD transfection reagent (Promega, USA) in the Opti-MEM Reduced Serum Medium (Life Technologies, USA) was added to cells in Grace medium. The cells were cultured for 72 h at 28°C before the promoter activity assay.

Genomic search for silkworm G-quadruplex and i-motif

BmPOUM2 orthologues in the six insect species were identified by blastp against NCBI nr database (http://www.ncbi.nlm.nih.gov/) and their upstream ∼500 bp were retrieved from their genomic sequences in NCBI GenBank (http://www.ncbi.nlm.nih.gov/). The accession numbers of the orthologues are XP_013185201.1 (Amyelois transitella), EHJ67329.1 (Danaus plexippus), XP_013178065.1 (Papilio xuthus), ACN43331.1 (Tribolium castaneum) XP_393686.2 (Apis mellifera), NP_001163355.2 (Drosophila melanogaster). Sequences were aligned using ClustalW package in Bioedit (Version 7.2.5) and manually annotated.

Genomic sequences, geneset and related annotation data of the silkworm were obtained from SilkDB (http://silkworm.swu.edu.cn/). Motifs of G4 (G3+N1–7G3+N1–7G3+N1–7G3+) and the corresponding i-motif were searched in the silkworm genome, respectively, using the program Quadparser (35). These motifs were then localized to specific genomic annotated regions with ≥ 1 bp overlapping cutoff.

RESULTS

Identification of the secondary structures in the promoter of BmPOUM2

To clone the promoter sequence of the transcription factor gene BmPOUM2, PCR amplification was performed using the primers designed based on the B. mori genome sequence deposited in SilkDB with genomic DNA isolated from B. mori larvae as templates. However, all trials for cloning the target sequence were failed under the general PCR condition. When 5% or 8% DMSO was added to the PCR reaction mixture under the same reaction conditions, the target promoter sequence (1638 bp) was obtained (Figure 1A). This result implied that the target promoter region may contain DNA secondary structures that block the Taq DNA polymerase binding with the DNA (36,37).

Figure 1.

PCR amplification of the promoter sequence of BmPOUM2 at different reaction conditions. (A) PCR reaction in the presence of 0, 5 and 8% DMSO. The faster migrating small bands are the non-target PCR products, non-specifically amplified. (B) The DMSO dose-dependent PCR assay with 0–10% (v/v) of DMSO in the PCR reaction mix. (C) The effect of denaturation temperatures on the amplification of PCR product without DMSO. (D) The effect of denaturation time on the amplification of PCR product. The denaturation temperatures kept at 94 or 95°C and the denaturation time were kept for 0.5 or 1.0 min. The general PCR conditions are described in Materials and Methods. The arrow shows the targeted PCR product.

For further investigation, the 1638 bp promoter sequence was cloned into pMD-18T vector, generating a 1638-pMD-18T recombinant plasmid. Because the G4 and i-motif are always found in the GC-rich region of DNA, the PCR primers were designed to amplify the GC-rich region only in the subsequent DMSO dose-dependent PCR. When the concentration of DMSO was increased higher than 3%, the target sequence (384 bp) started to be amplified (Figure 1B). When the denaturation temperature of PCR reaction was risen to 96°C or 97°C, the target sequence was also amplified (Figure 1C). When the denaturation temperature was maintained at 95°C but the denaturation time was extended to 60 s, the target sequence could also be amplified (Figure 1D). All these results implied that the BmPOUM2 promoter may contain secondary structures that can be deformed either by chemical (DMSO) or physical (higher denaturation temperature or longer denaturation time) factors.

Location of the secondary structures

Sequence analysis of the cloned promoter of BmPOUM2 indicated that it has a GC-rich region between –237 and +1 nucleotides (nts) (Figure 2A). To detect the accurate position of the secondary structures, the G-rich region of the forward strand of the promoter was truncated into three parts: –78 to +1 nts; –79 to –157 nts and –158 to –237 nts and subjected to CD analysis. The results indicated that the –78 to +1 and –158 to –237 regions had maximum and minimum absorption peaks at 280 and 250 nm, respectively, whereas the –79 to –157 region had maximum and minimum absorption peaks at 270 and 240 nm, respectively, and the maximum peak at 270 nm was enhanced when 100 mM K⁺, which stabilizes the G4 structure (38), was added to the DNA sample (Figure 2B). This CD spectrum change was well consistent with the reported pattern of absorption peaks of reported G4 structure (24,29,39).

Figure 2.

Identification of the G4 and i-motif structures in the promoter region of the BmPOUM2 gene. (A) The nucleotide sequence of the promoter region of the BmPOUM2 gene. The G and C bases are shown in red. (B) CD analysis of the +1 to –237 nts region of the promoter truncated into three fragments with or without 100 mM KCl. (C) Comparison of the G4 region of the BmPOUM2 promoter and some reported G4-containing sequences predicted by the QGRS Mapper. (D) CD analysis of the forward ssDNA of the –88 to –127 nts region of the promoter in absence or presence of 100 mM KCl. The mutated nucleotides are shown in the mutant sequence. F-WT: forward wild type strand; F-Mut: forward mutated strand. (E) CD analysis of the reverse ssDNA of the –88 to –127 nts region of the promoter at pH 4.1 and 8.0. The mutated nucleotides are shown in the mutant sequence. R-WT: reverse wild type strand; R-Mut: reverse mutated strand.

QGRS Mapper (http://bioinformatics.ramapo.edu/QGRS/) is a program that detects the composition and distribution of putative quadruplex forming G-rich sequences (QGRS) in nucleotide sequences (40). Analysis on the –79 to –157 nts region with QGRS Mapper revealed that the sequence of –94 to –120 region (GGGGCGCGAGGGGGCCGAGGGGCGGGG) has a higher G-score than many reported G4 sequences (39,41–48) (Figure 2C). The forward (F) and reverse (R) single strand DNA (ssDNA) of –94 to –120 nts were examined by CD for G4 and i-motif structures, respectively. For the F-strand, maximum and minimum peaks were observed at 265 and 240 nm, respectively, and the peak at 265 nm was increased when 100 mM K⁺ was added (Figure 2D), showing a typical feature of G4 structure (24,29,39). When the guanine at –95, –97, –100, –101, –108, –110, –118 and –119 sites were mutated to T, A, A, T, A, T, T and A, respectively, the maximum and minimum absorption peaks were shifted to longer wavelengths and the K⁺-mediated enhancement of the maximum intensity was suppressed (Figure 2D). For the R-strand, maximum and minimum peaks were observed at 290 and 260 nm, respectively, when the ssDNA solution was kept in acidic condition (pH < 6.0). However, the maximum and minimum peaks were shifted to 280 and 250 nm, respectively, at pH ≥6.6 (Figure 2E and Supplementary Figure S1), which is nearly identical to previous analyses of i-motif structures by CD spectroscopy (29,49,50). When the cytosine at –95, –97, –100, –101, –108, –110, –118 and –119 sites were mutated to A, T, T, A, T, A, A and T, respectively, the maximum and minimum absorption peaks were shifted to shorter wavelengths and the H⁺-induced enhancement of the maximum intensity was blocked (Figure 2E).

All these results of PCR, bioinformatic prediction and CD analysis suggested that a G4 and an i-motif structure were present in the F- and R-strand, respectively, of the –94 to –120 nts region of the BmPOUM2 promoter.

Effects of the i-motif structure on the transcription

To test whether or not the secondary structures are involved in the regulation of BmPOUM2 transcription, luciferase expression activity assay was performed. The wild-type or mutated promoter sequence (+72 to –1360 nts) was cloned into an expression vector (pGL3) to control the luciferase expression and the recombinant vector was used to transfect Bm12 cells. The results indicated that the mutated structure significantly decreased the activity (∼75%) of the promoter as compared to the wild-type structure (Figure 3A), suggesting that the secondary structure, either G4 or i-motif, or both, is necessary for the promoter activity.

Figure 3.

Effects of mutation and ASOs on the promoter activity of BmPOUM2. (A) Determination of the activity of the BmPOUM2 wild-type or mutated promoter by the luciferase assay. WT, wild-type; Mut, mutant; EV, empty vector. (B) Descriptive diagram of the principle of ASO assay. In normal conditions, the dsDNA was unwound to ssDNA and the G-rich and C-rich ssDNA fold to form G4 and i-motif structures, respectively (a); when the excessive ASOs of the G4 or i-motif were added, the G4 or i-motif would be inhibited (b). (C) Effect of ASOs interfering assay on the BmPOUM2 promoter activity. Non-specific-ssDNA: non-specific single-stranded DNA was used as a control. WT-ASO-i-motif and Mut-ASO-i-motif: wild-type and mutated complementary anti-sense oligonucleotides of i-motif. WT-ASO-G4 and Mut-ASO-G4: wild-type and mutated complementary anti-sense oligonucleotides of G4. F-ASO-Bmβ-actin and R-ASO-Bmβ-actin: complementary anti-sense oligonucleotides of reverse and forward strand. Data is means ± SEM (n = 3). ***P < 0.001 (Student's t test).

Addition of excess amount of ASOs complementary to a secondary structure sequence could inhibit its formation, allowing formation of normal dsDNA and resulting in inhibition or activation of transcription (24,51). This assay is not involved in any base mutation and can maintain one (either G4 or i-motif) structure while suppress the other, allowing to determine which structure contributes the regulation function (Figure 3B). The results of the ASO assay indicated that when the wild-type ssDNA that is complementary to the i-motif sequence (ASO-i-motif) was used to treat the cells, the promoter activity was significantly decreased (Figure 3C), implying the i-motif was suppressed by the excess amount of ASO-i-motif. The mutated ASO-i-motif did not have this inhibitory effect. When either wild-type or mutated ssDNA that are complementary to the G4 structure sequence (ASO-G4) was added, the promoter activity did not significantly change (Figure 3C). As negative control, complementary ssDNA fragments of the β-actin promoter (ASO-β-actin) did not change the activity because it does not contain any secondary structures.

The results of ASO assay indicated that only the i-motif structure participated in the regulation of BmPOUM2. Therefore, the regulatory function of i-motif structure was only focused in the subsequent researches. The porphyrin compound TMPyP4 (Figure 4A) has been reported to be able to bind with the G4 and i-motif structures (52,53). The TMPyP4 isoform, TMPyP2, cannot bind with G4 and i-motif structures and can be used as negative control (29,54). To determine whether the BmPOUM2 i-motif can interact with TMPyP4, CD analysis was performed. The results indicated that TMPyP4, compared to TMPyP2, could dramatically deform the i-motif structure when the concentration reached 25 μM. When the concentration of TMPyP4 exceeded 50 μM the i-motif almost disappeared (Figure 4B). In addition, the CD analysis of melting temperature indicated that the TMPyP4 could dramatically decrease the melting temperature (T_m from 62 to 46°C) of the ssDNA fragment that contained the i-motif structure (Figure 4C). TMPyP2 did not have this effect. These results implied that the BmPOUM2 i-motif structure can be deformed by TMPyP4.

Figure 4.

Effects of TMPyP4 and TMPyP2 on the i-motif structure and promoter activity of BmPOUM2. (A) The structure of the porphyrin compounds TMPyP2 and TMPyP4. (B) CD analysis of the i-motif structure in the presence of TMPyP4 or TMPyP2. (C) CD analysis of melting temperature of the i-motif structure in the presence of TMPyP4 or TMPyP2. The synthesized ssDNA oligonucleotides that contained the i-motif region (5 μM) were heated at 95°C for 10 min in 50 mM Tris-acetate buffer at pH 4.1 and slowly cooled to room temperature over 4 h period to allow i-motif structure to form. TMPyP2 or TMPyP4 was then added into the solution at a final concentration of 25 μM and incubated overnight at 4°C, followed by CD analysis. The dot lines show the melting temperatures. (D) Determination of the promoter activity by the luciferase assay in the presence of TMPyP4 or TMPyP2. Data are means ± SEM (n = 3). ***P < 0.001 (Student's t test).

If this promoter sequence can form i-motif structure in cells, the promoter activity of BmPOUM2 should be changed when the cells were treated with TMPyP4. In the promoter activity assay, TMPyP4 or TMPyP2 was used to treat the cells when the recombinant vectors were transfected in cells. The results showed that comparing to negative control TMPyP2, TMPyP4 significantly suppressed the wild-type promoter activity (Figure 4D). However, when the nucleotides at -95, -97, -100, -101, -108, -110, -118 and -119 sites were mutated to A, T, T, A, T, A, A and T, respectively, the inhibition effect of TMPyP4 disappeared (Figure 4D), suggesting that TMPyP4 interacted with the wild-type structure of i-motif and impacted the transcription function of the promoter.

These results demonstrated that when the i-motif structure was deformed by site mutation, ASOs or TMPyP4, the promoter activity of BmPOUM2 was decreased, revealing that the i-motif structure is involved in the regulation of the BmPOUM2 transcription.

Identification of nuclear protein that binds to the i-motif structure

To identify the nuclear protein that binds to the i-motif structure and regulates the transcription of BmPOUM2, pull-down experiments were conducted. The nucleotide fragment of the i-motif structure was labeled with biotin and then linked to the streptavidin-coated beads. The complex was then incubated with the nuclear proteins isolated from Bm12 cells. The proteins that bound to the biotin-labeled i-motif probe were separated on a SDS-PAGE gel and the differentiated protein bands were excised and subjected to LC–MS/MS analysis. The results showed that there was one differentiated protein band for the i-motif probe, compared to the mutated probe (which could not form i-motif structure). The size of the protein band for the i-motif probe was approximate 41.5 kDa (Figure 5A). While the identification of the G4 binding proteins still is ongoing, a homolog of the human interleukin enhancer binding factor (ILF, GenBank accession no.: NM_001046918.1) (55) was identified from the candidate i-motif binding proteins (Supplementary Table S1). To determine whether this protein binds to the i-motif, the ORF of the B. mori ILF (BmILF, GenBank accession no.: KY082711) was cloned. A recombinant protein was expressed in E. coli and purified for EMSA. The EMSA assay showed that the recombinant BmILF specifically bound to the labeled i-motif probe and the binding was enhanced with the increase of the protein concentration and competed off by the unlabeled probe (Figure 5B). The BmILF did not bind to the mutated probe, which could not form the i-motif. To further demonstrate that the BmILF was binding to the i-motif secondary structure rather than to the linear ssDNA, the concentration of native-PAGE was increased from 6% to 12%. In 12% native-PAGE the patterns of the i-motif and linear ssDNA were distinguished (Figure 5C). In the absence of BmILF, the DNA probe at pH 4.1/6.0 was separated into two different bands, but at pH 8.0, only one band was found (Figure 5C). So, the lower bands were the i-motif structure and the higher bands were linear ssDNAs. When BmILF was added to the reaction system, the i-motif probe band was decreased and a new higher band (BmILF+i-motif probe) was appeared. The linear ssDNA band was not affected (Figure 5C). This result indicated that BmILF did bind to the i-motif structure, rather than to the linear ssDNA.

Figure 5.

Identification of nuclear proteins that bind to the i-motif structure. (A) Pull-down experiment with nuclear proteins of Bm12 cells. The oligonucleotide probes were shown in the top panel. WT and Mut: wild-type and mutated ssDNA. The arrows point to the protein band observed in WT but not in the mutant. (B) EMSA (6% PAGE) analysis of the recombinant BmILF binding with the i-motif probe. The cold probe was the unlabeled i-motif probe. (C) EMSA (12% PAGE) analysis of the recombinant BmILF binding with the i-motif structure or linear DNA probe at different pH conditions. The positions of the labeled i-motif-containing probe, labeled linear DNA probe and the labeled bound i-motif and BmILF are shown by arrows. (D) Bm12 cells transfected with EGFP vector as a control or with BmILF-EGFP. The green fluorescence shows the similar transfection efficiency and expression of EGFP and BmILF-EGFP. Chromatin immunoprecipitated (ChIP) target sequence was detected by RT-PCR (top panel) and by qRT-PCR (bottom panel). The enrichment of the promoter sequence in immunoprecipitated DNA samples was normalized with DNA present in the 10% input material. Data is means ± SEM (n = 3). ***P < 0.001 (Student's t test). (E) The reverse strand sequence of the -68∼-153 nt region of the BmPOUM2 promoter. The region that forms the i-motif structure is boxed in red. The primer aligned regions are underlined. (F) The sequencing atlas of the enriched RT-PCR product of the ChIP assay.

To further demonstrate whether or not BmILF can bind to the i-motif of the BmPOUM2 promoter ex vivo, ChIP assay was performed. An EGFP tag was fused to BmILF and anti-EGFP tag antibody was used in the ChIP assays (56). The cells were transfected with the recombinant plasmid BmILF-EGFP or EGFP (as a control) (Figure 5D) and then collected for ChIP assay after 48 h post transfection. The results showed that the anti-EGFP antibodies, but not IgG (as a negative control), precipitated i-motif fragment of the BmPOUM2 promoter in the cells transfected with the BmILF-EGFP expressing plasmid (Figure 5D). The enriched i-motif sequence amplified by PCR was confirmed by sequencing (Figure 5E and F). This experiment demonstrated that BmILF did bind to the i-motif of the BmPOUM2 promoter ex vivo.

The regulatory function of BmILF

To determine whether or not BmILF influences the BmPOUM2 promoter activity, the BmILF-EGFP expression construct and the pGL3 BmPOUM2 promoter-luciferase construct were co-transfected into Bm12 cells and the reporter activity was quantified after 48 h post transfection (Figure 6A). The BmPOUM2 promoter-controlled luciferase activity was increased significantly in the cells transfected with BmILF-EGFP, but not in the control cells transfected with EGFP (Figure 6B). Complementary to the over-expression experiment, BmILF RNAi was performed. When the BmILF expression was down-regulated by dsBmILF in Bm12 cells (Figure 6C), the activity of the BmPOUM2 promoter was significantly decreased (Figure 6D).

Figure 6.

Regulation of the BmPOUM2 promoter activity by over-expression or RNAi suppression of BmILF. (A) Bm12 cells co-transfected with EGFP vector as a control or with BmILF-EGFP. The green fluorescence shows the similar transfection efficiency and expression of EGFP and BmILF-EGFP. (B) The luciferase activity in Bm12 cells co-transfected with EGFP vector or BmILF-EGFP and the pGL3-WT-BmPOUM2 promoter-luciferase vector (WT-BmPOUM2-P-Luc) or the i-motif-mutated pGL3-Mut-BmPOUM2 promoter-luciferase vector (Mut-BmPOUM2-P-Luc). (C) Reduction of BmILF by RNAi detected at 48 h and 72 h after BmILF dsRNA was transfected into Bm12 cells. (D) Changes of the luciferase activity in Bm12 cells co-transfected with BmILF dsRNA or EGFP dsRNA and the luciferase reporter vector. Data in (B) and (D) are means ± SEM (n = 3). **P < 0.01 (Student's t test).

All the experimental results of the up-regulated and down-regulated BmILF expression suggested that the nuclear protein BmILF bound to the i-motif structure in the BmPOUM2 promoter and regulated its transcription.

DISCUSSION

G4 structure widely exists in human genome (57,58) and it has been found to exist in the regulatory regions of genes and distribute near the transcription start sites (TSSs) (19,20), implying that G4 secondary structure may function as an important regulatory cis-element. The C-rich complementary strand of G4 can form i-motif, but relatively less studies on the i-motif function have been reported. In this study, G4 and i-motif structures were identified in the GC-rich regions (between –120 and –94) of the BmPOUM2 promoter, which can be deformed either by chemical (DMSO) or physical (higher denaturation temperature or longer denaturation time) factors. This is the first report on the secondary structures present in the regulatory region of a transcription factor gene in lepidopteran insects.

To examine whether or not these secondary structures are also present in the POU homologue genes and the silkworm genome, the promoter sequences of POU homologue genes of six species were analyzed. Five POUs promoter showed predicted G4 and i-motif structures (Supplementary Figure S2). In the B. mori genome, 3174 G4 and 3104 i-motif structures were found with 357 motifs (147 G4 and 210 i-motif) being in the upstream region of 322 genes (2.27% of total gene set) (Supplementary Table S2). These results imply that DNA secondary structures are widely and commonly distributed in the genome and may play important roles in the regulation of gene transcription in insect.

Studies on the transcriptional roles of G4 and i-motif reveal that these structures play inhibitory or promoting effects on gene transcription. In some cases, they may have repressive effect. For example, in WNT1, c-MYC (21–23), the G4 structure suppresses the gene expression, probably due to prevention of the indispensable transcription factors from binding to the promoter or due to facilitation of binding of transcription suppressors to the promoter. In other cases, the secondary structures have activation effect, for example, in col2al, fdz5, nog3 (24) and muscle specific gene promoter (59), probably due to recruitment of a transcription activator or due to estrangement of transcription suppressors. In this study, it is interesting to find that when the i-motif structure was removed or deformed by either mutation, ASOs, or TMPyP4, the promoter activity was decreased significantly (Figures 3 and 4), suggesting that the i-motif structure in the BmPOUM2 promoter benefits or enhances its expression.

The most important finding of this study is that a novel i-motif bound protein, BmILF, was identified and demonstrated by using the methods of DNA-protein pull-down, EMSA and ChIP assays. Although many G4 binding proteins have been reported (60), only one i-motif binding protein (isoform 1 of heterogeneous nuclear ribonucleoprotein L-like, hnRNP LL) was reported so far in detail (31,61,62). hnRNP LL contains four RRM (RNA recognition motif) domains that were predicted to bind with i-motif, resulting in activation of the bcl-2 gene expression (31,61). BmILF does not have any RRM domain. In human, ILFs have three isoforms, ILF-1 (alias FOXK2), ILF-2 (alias NF45) and ILF-3 (alias NF90), and are involved in regulation of different processes of development and metabolism. The ILF-1 belongs to forkhead gene family (63). The ILF2 and ILF3 both contain a domain associated with zinc fingers (DZF). BmILF also contains a DZF domain and shares about 53% amino acid identity with human ILF2. The human ILF2 and ILF3 can dimerize and act as coactivators to initiate the expression of the c-fos gene by association indirectly with RNAPII (RNA polymerase II) (64). However, no study on ILF binding to the i-motif or G4 structure has been reported so far. BmILF is the first identified novel protein that contains a DZF domain and binds with the i-motif structure to activate the expression of BmPOUM2.

A regulatory model for the interaction between the i-motif structure and BmILF is suggested based on the data presented in this study (Figure 7). When the transcription of BmPOUM2 is initiated, the duplex DNA strands separate and basal transcriptional factors assemble. The single strand DNA in the GC-rich region forms G4 or i-motif structure due to negative supercoiling. The formed i-motif structure then recruit its binding proteins required for the transcription to activate the transcription complex, turning on the transcription of BmPOUM2. After transcription, the basal transcription factor(s) disassemble from the core promoter and the secondary structures deformed to be dsDNA, turning off the transcription.

Figure 7.

Diagram of proposed epigenetic regulatory mechanism of BmPOUM2 by the i-motif secondary structure, which was bound by BmILF. When BmPOUM2 starts to transcribe, the basal transcription factors bind to the core promoter, separating the dsDNA strands. At this time, the GC-rich dsDNA in the upstream region would open because of negative supercoil. The G-rich and C-rich ssDNA immediately fold to form G4 and i-motif structures, respectively. The i-motif recruits the transcription factor BmILF, which may recruit other co-factors required for expression of BmPOUM2. After transcription, the basal transcription factor(s) disassemble from the core promoter and the secondary structures unfold to form dsDNA, turning the transcription off.

One important issue needs to be addressed. Can the i-motif structure actually form and function in in vivo physiological condition? This study demonstrated that the i-motif of BmPOUM2 could ex vivo form and bind with BmILF in the cells cultured in vitro, regulating the expression of BmPOUM2. DNA exists and gene transcription takes place in the nuclei of cells, which have an acidic environment in favor for the formation of i-motif structure. However, further investigation still needs to directly demonstrate the existence and regulatory function of the i-motif structure in vivo.

Regulation of gene (or protein) expression can take place at three levels: transcription, post-transcription and translation. The regulation can be genetic (such as gene and promoter primary sequences), epigenetic (for example, methylation and histone modification) or physiological (e.g. mRNA splicing, interference, ribosome assembly). All of these mechanisms involve in only the linear primary sequences with strict A–T and C–G base-pairings. The presence and function of the secondary structures in the DNA and RNA molecules provide a novel epigenetic mechanism for gene transcription regulation. Like methylation and histone modification, formation and deformation of these structures are not directly determined by genetic codes. This mechanism has several typical features. Firstly, it is epigenetic and does not depend on genetic codes of primary sequences. Secondly, the secondary structures stay in a dynamic balance between formation and deformation depending on the status of the cells, in other words, base-pairings can be changed between C–G and G–G/C–C+, which is reversible. Thirdly, the secondary structures can form within ssDNA or mRNA molecules and between two strands of two molecules of DNA or RNA (18). Fourthly, the formation and deformation of the secondary structures may need the involvement of proteins. Once either the formation or deformation of the secondary structures is disturbed by inhibiting the binding proteins, the gene transcription would be suppressed or activated (29,65). This provides a novel target and approach to controlling gene expression, cell function and new pesticides for pest control, gene therapy for diseases, independent of transgenic manipulation.

In summary, this study for the first time identified i-motif structure present in the promoter region of a lepidopteran transcription factor gene in B. mori. The i-motif structure, bound with the nuclear transcription factor BmILF, was involved in the transcription regulation of the gene. All these results together provide new insights into the molecular mechanism of the i-motif structure in epigenetic regulation of gene transcription. The mechanism of the in vivo formation and function of the i-motif structure in physiological conditions still needs to be further investigated, although some studies have provided some possible mechanisms (26,27).

AVAILABILITY

The amino acid sequence of the transcription factor BmILF is deposited in GenBank (GenBank accession no.: KY082711).

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online.

FUNDING

Chinese National Natural Science Foundation [31330071, 31672494, 31720103916, 2012CB114602]. Funding for open access charge: South China Normal University.

Conflict of interest statement. None declared.